mdhub

mdhub is an AI operating system for behavioral health clinics, built by mdhub, Inc. in San Francisco (founded 2023, YC S24, $500K seed round): Emma turns session recordings into clinical notes, Sarah answers the phones, and Eric handles billing. It does not crowd into the generic transcription market; what it sells is a full back office for psychiatric, psychotherapy and addiction clinics.

Core capabilities

• Emma notes: records the whole session and produces a structured note in about 30 seconds using SOAP/DAP/BIRP templates, including pre-charting and CPT/ICD-10 coding
• Sarah intake: answers calls, texts and faxes 24 hours a day, and automates insurance eligibility checks, screening and scheduling
• Eric billing: from coding through claim submission, denial follow-up and prior authorizations
• Chart-aware GPT: asks questions about the chart context inside a HIPAA environment and drafts follow-up summaries
• Group therapy splitting: splits a single recording across each patient's chart by speaker attribution

Where it fits Built for mid-size and larger US behavioral health organizations: psychiatric, psychotherapy and addiction treatment clinics. The customers shown on its website are multi-site organizations such as Family Care Center and Talkiatry. It suits teams that want one system to replace three separate backends for notes, intake and billing; solo practitioners who only want recording plus notes can skip it.

What sets it apart Competitors at the same price point are mostly single-purpose scribes, while mdhub assembles a set of "AI employees" into a full stack, and in 2026 it is building its own EHR/CRM/RCM as well. Its compliance coverage is only HIPAA and SOC 2, with no HDS and no CE/MDR, so it is a thoroughly American business.

Editorial take The convenience is real: users report that catching up on notes after hours dropped from an hour a day to nearly nothing, and that templates adapt to personal style. The weaknesses are just as real: it covers behavioral health only and does not touch imaging or labs; the interface is English only; it runs on Azure OpenAI underneath with no in-house medical terminology database; mobile reviews are split, with App Store at 4.2 from only 10 reviews and Google Play at 2.0 from only 6; and with a $500K seed round and a 25-person team, whether the vendor survives the next few years is a calculation you have to make yourself. Choosing it means choosing the US market.